﻿Replication Materials for “Veneration or Respect? A Multidimensional Framework for Understanding Public Attitudes Toward Constitutional Change”
Andrew Kim


OVERVIEW
This repository contains replication code and documentation for the various analyses reported in the manuscript and appendix. The original respondent-level survey data was collected with IRB approval under the restriction that all data were destroyed following analysis, thus it cannot be made publicly available. The repository nevertheless provides complete replication code and detailed variable documentation.
--------------------------------------------------
REPOSITORY CONTENTS


analysis.R - Single R script containing all data pre-processing, variable construction, descriptive analyses, factor analyses, statistical power checks, and regression models reported in the manuscript and appendix.


README.txt
This file.
--------------------------------------------------
SOFTWARE REQUIREMENTS


Analyses were conducted in R. Required packages include:


dplyr, magrittr, tidyr, forcats  
readxl, stringr  
psych, pwr  
ggplot2, scales, grid  
knitr, stargazer, jtools  


All packages are available on CRAN.
--------------------------------------------------
REPLICATION INSTRUCTIONS


The replication script assumes a raw survey file located at:


data/raw/survey.csv


The raw file should contain one row per respondent and original survey response variables as exported from the survey platform. Appendix 1 in the Supplementary Materials documents the full survey instrument employed.


Analytical methods in the replication script include:
- Data cleaning and recoding
- Construction of composite indices
- Factor analytic validation
- Descriptive statistics and visualizations
- Statistical power analysis
- OLS regression models with alternative specifications
- Robustness checks across outcomes


All tables, figures, and analysis results reported in the manuscript are generated by this script.
--------------------------------------------------
VARIABLE DOCUMENTATION


Variables are documented after pre-processing and feature construction, as used in final analyses. Unless otherwise noted, scale variables are rescaled to the [0,1] interval.


Constitutional Respect


- respect - mean of all statements in constitutional respect battery
- symbolic - mean of statements measuring symbolic respect dimension
- relevant - mean of statements measuring perceived modern relevance dimension
--------------------------------------------------
Amendment Rigidity


- rigidity - mean preferred rigidity threshold for steps in constitutional amendment process
- rig_congress - preferred rigidity threshold at congressional approval stage
- rig_states - preferred rigidity threshold at state ratification stage


Response coding:
0 = Current rigidity is too high
0.5 = Current rigidity is just right
1 = Current rigidity is too low
--------------------------------------------------
Amendment Support


- support - mean support for all hypothetical constitutional amendments presented in survey
- support_conserv - mean support for conservative-leaning amendments (i.e., flag desecration ban, abortion ban)
- support_liberal - support for liberal-leaning amendments (i.e., gender equality, gun control, - electoral college abolition)


Individual amendments presented include:
amend_flag, amend_abort, amend_term, amend_gender, amend_guns, amend_elect
--------------------------------------------------
Political Trust


- trust - mean trust in political institutions, incl. federal government, state government, courts, and political parties
--------------------------------------------------
Constitutional Knowledge


- knowledge - composite constitutional knowledge score based on correct selections/omissions to factual questions about U.S. Constitution
--------------------------------------------------
Respondent Demographics


- Race and ethnicity dummy indicators (0 = not member of group, 1 = member of group): white, black, asian, aian, nhpi, hisp, other_race_eth
- gender - binary indicator for gender (0 = male, 1 = female); non-binary resources coded as missing due to small sample size
- age - age calculated from year of birth
- income - categorical household income factor, collapsed to reduce multicollinearity (reference category: less than $10,000)
- citizen - dummy indicator for U.S. citizenship (0 = non-citizen, 1 = citizen)
- grad - dummy indicator for education level (0 = some college or lower, 1 = bachelor’s degree or higher)
- party - binary indicator for party affiliation (0 = Democratic, 1 = Republican), including partisan leaners